[][src]Trait dumbnet::activation::Activation

pub trait Activation: Debug + Serialize + DeserializeOwned + Clone {
    fn activate(inputs: f32) -> f32;
fn derivate(activation: f32) -> f32; }

Required methods

fn activate(inputs: f32) -> f32

Input is the summed and weighted inputs of this neuron

Output is the output of the neuron

fn derivate(activation: f32) -> f32

Input is the previous activation, as in the result of this traits activate() function

Output is the derivate of this activation function at that point. The derivate may be slightly adapted to better suit the needs of gradient descent

Loading content...

Implementors

impl Activation for ReLu[src]

impl Activation for Sigmoid[src]

Loading content...